Roman Sandals

March 27, 2008

Infrastructure Money

Filed under: musing, technology — Craig Lawton @ 2:33 pm

It’s interesting to note that in big IT shops software often accounts for about 50% of the operating budget. Compare this to human resource, or people as they are sometimes referred to, which comes in at about 10% of most budgets.

In the last downturn, server hardware got hammered in cost, hurting the likes of SUN Microsystems, and helping the likes of Dell. People got done what they could, with what they could budget for.

If the current economic storm clouds start hurting IT budgets, will software (and possibly also Storage costs) be the sacrificial lamb, opening doors for the likes of MySQL, Debian,  Tomcat, Nagios etc. to make big in-roads in to the corporate world? Could SUN make some inroads with ZFS (over VxFS)?

February 14, 2008

A basic systems testing toolkit

Filed under: sysadmin, test-driven sysadmin — rchanter @ 9:41 am

Looking around for examples of test-driven sysadmin, all I can find is people recommending, rather sensibly, that you test systems changes before deploying them to production. I’m interested in both a broader and more narrow view of a systems testing toolkit.

Broader in the sense that I want to test many more things than just planned change. If we’re talking about a test-driven approach (and, ultimately, a behaviour-driven approach), then we should apply a testing mindset to all the activities of systems administration; incident management, problem management, change management. More specific repeated activities like simple health-checks, verifying the correctness of data changes (as opposed to configuration changes), and so on.

Narrower in the sense that I want a simple, flexible toolkit that lets me express tests with a common language and collect results in a common format.

In all of this, simplicity and ubiquity are key considerations. They influence choice of language (and even of coding style), choice of file formats, and infrastructure design (hint: there isn’t any). I’ll go into more detail in my next post.

Anyway, developers have long had unit testing toolkits available to them: JUnit, Test::Harness, RSpec, the list goes on. While none of them are a great fit for systems testing, there is plenty of inspiration we can take from looking at them.

February 5, 2008

He’s Baack

Filed under: inconsequentia — rchanter @ 7:39 am

Penny Arcade – FF2K @ CES2008 (mildly NSFW).

January 17, 2008

Save the kiddies from!!!

Filed under: political — Craig Lawton @ 3:52 pm

The new government’s proposed “clean feed” net filter system mandates that ISPs block sites from a known “bodgy” list.

The filtering infrastructure will be a bottleneck, slowing down and making more expensive our Internet connectivity. And the ALP went to the election promising to improve our broadband access and speeds!!!

The system will not stop paedophiles preying on children in chat rooms, nor stop kiddie porn. That’s not what it does.

The system will give the illusion of protection, which is probably more dangerous, because it may stop parents taking more direct action to control Internet access.

I can imagine the blacklist will get distributed some day somehow, so adults who fancy pornography can browse the government’s “recommendations”. I guess children may get this list too but I can’t imagine they’d be interested in looking at a bunch of sites their parents don’t want them to look at!!

Check out the Senate results from last year’s election and you get the story. For the ALP to get legislation through the Senate (assuming the Coalition votes against), it needs the support of the Greens. But that is still not enough. The legislation needs one more vote from either Nick Xenophon or Family First’s Steve Fielding. It looks like the ALP has adopted Family First policy as a future favour-saver!

Sign the petition here:

January 15, 2008

Thinking About Test-Driven Systems Administration

Filed under: sysadmin, test-driven sysadmin — Tags: , — rchanter @ 1:59 pm

So I’ve been thinking about, and working on, this for a little while.

It was prompted mainly by the title of a paper Geoff Halprin gave at last year’s SAGE-AU conference. Not having had the time to attend the conference itself last year, I have no idea whether my approach bears any resemblance at all to Geoff’s.

Broadly speaking, systems administration consists of two main tasks: managing planned change to systems, and managing unplanned incidents on systems. Everything else we do is just arranging affairs so that change is simpler and more deterministic, and incidents are shorter and less frequent.

How, then, can a test-driven approach help with this? It seems to me that we need two things:

  • A test-first workflow that makes sense for systems management
  • A language and toolkit for expressing tests and collecting the output.

This seems like a pretty simple task. I’ll explore it a bit in the next few weeks.

Neat systems management hack using clamav

Filed under: sysadmin — Tags: , — rchanter @ 1:56 pm

From the clamav-users list, last year (yes, I’m behind on my reading):

Clamav was used in Debian to discover copies of statically linked
copies of zlib that needed a security update.

December 11, 2007

This is why we invented the Internet

Filed under: inconsequentia — rchanter @ 10:28 am

Go watch a few episodes of this, then try and tell me I’m wrong.


Filed under: spam, technology, Uncategorized — rchanter @ 10:27 am

So one of our mail servers got listed on spamcop the other day. It’s just an operational hazard of running a mail service of a non-tirvial size really, but still a PITA. Delisting is simple enough, mopping up is harder. I don’t know who I should be most annoyed with:

  1. Spamcop, for being a trigger-happy, FP-prone list (and by extension, Ironport for not doing enough to clean up their act).
  2. The people running mail servers who think spamcop is a safe RBL. This includes a few providers that I would have expected to know better.
  3. The people running rogue autoresponders inside our network, which is the most likely way for reputable senders to hit the spamcop spamtraps.
  4. IBM/Lotus, whose Out-of-Office autoresponder is an utterly brain-dead piece of crap. (and don’t get me started on how unusable mail rules are).

The right answer, of course, is all (or none) in equal measure. But deep down, I think I want to blame Spamcop and Ironport. Now, I’m all for blacklists discouraging backscatter. But no matter what measures the service operator takes, there’s always going to be something back at the mailbox that does The Wrong Thing. And Spamcop (by which I mean Ironport) have a tool that would be exactly the right thing to help distinguish between indiscriminate backscatterers and sites that mostly have the problem under control.Grrr. B’stards.

November 22, 2007

C: Drive

Filed under: musing, technology — Craig Lawton @ 4:08 pm

I just had a thought. I should really back-up my work laptop. I should back-up my C: drive. I fired up the XP Backup tool.

Even plebs know what a C: drive is! But why is it so? Surely, it should be the A: drive if it’s that important. But no, the A: drive was originally assigned for a floppy disk, and I think the B: drive was for a secondary floppy disk from memory. But nobody has floppy disk drives anymore!

As if reading my mind, the XP Backup tool told me that after creating the back-up file it would ask me for a floppy to create a boot disk. Of course, you’d think it’d check to see if I had a floppy drive before asking. I don’t.

Interestingly, Wikipedia lists all the Operating Systems that use drive letter assignment. It reads like one of those Human Rights Watch charts, one which lists the countries that kill more of their own citizens than others: (more…)

November 15, 2007

Random Numbers and Pokies

Filed under: social, technology — Craig Lawton @ 2:09 pm

Australians love pokies. The Productivity Commission found in 1999 that we have 21% of the world’s machines in our pubs and clubs.

Somewhere in the dim, distant past poker machines became very much like computers. Somewhere even dimmer and further back in time, I was taught that computers were really bodgy at coming up with random numbers. Indeed they can only come up with pseudo-random numbers.

Gambling venues today prefer the decidedly “pseudo-name” of gaming venues, and maybe for good reason. A device that can guarantee a particular financial return, typically set at 75%-85%, isn’t taking chances. There is negligible risk to the house.
It doesn’t sound random to me. It doesn’t sound like gambling.

Some people may call it entertainment; some people may call it robbery.

« Newer PostsOlder Posts »

Create a free website or blog at